[Gluster-users] local mounts failing using systemd during bootstrapping gluster

[hvjunk]

Good day,

The specific SystemD problem reference is  https://github.com/systemd/systemd/issues/4468#issuecomment-255711912 <https://github.com/systemd/systemd/issues/4468#issuecomment-296904675>

This problem with GlusterFS arises specifically during bootstrapping of the cluster, ie. configure 3 VMs and the volume (2replicas and the arbiter) test it working, do local mount and setting it up in /etc/fstab. Once everything works, shutdown all the VMs in the cluster, and then start all of them together.

The expected result would’ve been that the /etc/fstab local fuse mount will all be mounted correctly, but only one or perhaps two, if at all, gets mounted. the problem being that the mount “failed”, doesn’t retry etc.

The hack around it, I used a SystemD service that , as a pre-exec mount the gluster volume through the systemd .mount mechanism, but then checks a directory inside the expected mount, and will then fail if that mount doesn't exists.

So, how do others solve this bootstrap mounting problem of GlusterFS local mounts on SystemD systems?

HEndrik


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.gluster.org/pipermail/gluster-users/attachments/20170502/68b9bf18/attachment.html>